t tests


t tests (like one sample, related samples, and independent) are just testing the mean of the samples means against the population mean, correct? Also, if I were to do a related t test but found out beforehand that the correlation between the two datas was not significant at .05 level, wouldn't it makes sense not to do a t test? Thank you!


TS Contributor
The one-sample t-test compares the mean of a sample with that of a standard value or a population value. The related samples test compares two related samples (observations are paired across the samples, and you compute a delta). The independent samples test compares two independent samples.

Related samples do not need to have a strong mathematical correlation in order to be called "realted" or "dependent" - all that is required is that they are not sampled independently of each other.


no, please attend:
if mu0 be mean of population and mu be mean of sample so :
one samle t-test is for testing this hypothesis: mu=mu0 i.e comparing sample mean with population mean
two saple t- test (paired or independent) for testing this hypothesis:
mu1=mu2 i.e comparing two sample mean

excuse me but you have not understand meaning of paired(related)sample
whatever you meant it ,it is not true.